基于改进Yolov5s的增强现实文物识别方法OACSTPCD

Applying augmented reality technology to the cultural relics exhibition of the digital museum helps to shorten the distance between the visitors and the cultural relics,and make the exhibition more interesting.Aiming at the problems of false detection and low recognition accuracy caused by complex background and rich shape and texture of cultural relics collected by augmented reality equipment in the application scene of cultural relics exhibition,an improved cultural relics recognition method based on Yolov5s is proposed.The CBAM attention mechanism is introduced between the backbone network and the neck network in the Yolov5s network structure.In addition,a multi-head attention mechanism is used to replace common convolution in the Bottleneck module of the backbone network,which effectively captures local information and reduces the interference of unnecessary information.In order to improve the location accuracy of the boundary frame of the target cultural relics,the DIoU-NMS method is used to select the optimal target identification frame as the final prediction frame.Experimental results show that the proposed algorithm improves the average detection accuracy of the cultural relics,which is more suitable for the target detection of cultural relics than the original algorithm in AR application.